The stochastic process model, as a powerful mathematical tool, can simulate and predict stochastic phenomena over time. The study adopts the Markov process prediction model in a stochastic process and incorporates the gray prediction model to construct a gray Markov model to predict the emergency response time for public emergencies. The performance of the model’s prediction is evaluated by comparing its accuracy to current mainstream prediction methods. The model is used to predict the emergency response time by simulating the water pollution accident in the Huaihe River section in Anhui Province. The model predicted that the response time of each water plant pollution accident during the dry and abundant water periods was less than the time when the pollutants reached the highest concentration, indicating that the emergency response time of public emergencies predicted based on the improved Markov process model was more adequate.